Blockchain. Disruption or Seduction?
Is blockchain disrupting businesses or merely seducing us with mathematical beauty and utopian dreams?
If you are puzzled by how blockchain can be a game-changer, you are not alone. Unlike some in the industry would claim, it may not be because you do not understand it.
When I first learned of Bitcoin and the blockchain I was wowed and impressed by the cleverness and application of abstract mathematics to real problems. But as I examined the use cases and claims, it appears there is little to justify the claims that blockchain is a “disrupter”, “game changer”, or “internet 3.0”.
The danger from being seduced by blockchain is that less costly and more efficient solutions to the problems will be ignored. Furthermore, funding gets diverted from other meaningful projects.
Finally the blockchain hype promotes investments in cryptocurrencies, which will pay off a technology elite and leave others holding the bag. This problem is due to two misconceptions that appear in media write ups on blockchain. The first is that greater usage of any platform anointed as a blockchain implies cryptocurrencies are becoming successful. The second is conflating any blue-sky investigation of a cryptocurrency platform with true successful adoption.
In this piece I will examine some of the use cases claimed, and why blockchain only provides the illusion of a solution.
What blockchain Does — Trust and Validation
Blockchain is sold as a disruptive solution to the “trust” problem, the requirement that people and businesses need to trust others in a transaction. Careful examination shows that it is often a costly solution that solves only a narrow sliver of the trust problem.
What blockchain offers is merely the ability to readily detect alterations in data that was registered in the chain. My omission of the decentralization and replication is deliberate. In some blockchain solutions the only data element that is stored on the blockchain is a fingerprint (“hash”) of the document or data of interest; one still must rely on a separate system, such as another centralized service or a separate blockchain, to store the data.
Whether the blockchain stores the data or the hash of the data, a blockchain can make it next to impossible to alter the data without being detected. This is because corrupting any fingerprint stored in the chain to match the altered data would make it inconsistent with the later records or blocks in the chain, each of which includes a fingerprint of its prior block. Thus corrupting any record is as hard as recreating the entire blockchain after that record. This is the feature that blockchain advocates generates the “trust” that is their concern.
Technologies developed in the late 20th century, which are employed in blockchains, can insure a similar degree of data immutability. These technologies include digital signatures and Merkle trees. In fact, because it is such a popular buzzword, systems using such technologies are marketed as blockchain solutions, even if there is no “blockchain.” Such systems may be excellent solutions, but not because they are blockchains.
Blockchain technology itself does nothing to assure us that the data that is linked in the chain was valid when it was added. GIGO — -Garbage In Garbage Out still rules the day. For most of the use cases of blockchains, this is the greatest underlying problem. The blockchain cannot prove the existence of a physical asset; it only proves that someone (or a bot) possessing the appropriate password[1] says that the asset exists and meets the stated specifications and that he or she made this claim prior to the next block of data being entered onto the chain.
I have been involved with many issues involving bad data. They all involved failures in entering the data into the system or logic used in processing — none addressed by blockchain.
What a blockchain does — Replication
The other common advantage cited for blockchains is their resiliency; in some blockchain solutions many copies of the data are made across blockchain nodes. Thus a hacker would have to attack all nodes to destroy the data.
The problem with this claim is this resilience is not unique to blockchains. Mirroring and other techniques are used to protect high value data currently. In cases where people are not using available technologies, why would they use a blockchain?
What can make a blockchain achieve greater resilience is having many more replicas than the owners of the data would choose to pay for in a more centralized solution. This creates a performance and network bandwidth cost. Of course, many commercial blockchain solutions will solve this by having a limited number of nodes replicate the data. Although this may be a good solution for some problems, it is hardly a game changer.
As I show in the examples below, many of the use cases of blockchains can and are being satisfied by alternatives, and in many cases these alternatives are working well.
Provenance and Supply Chains
A popular use case for blockchain is tracking goods through supply chains so that one can be assured of proper handling or compliance with good social and environmental conduct. This is not a function that necessarily requires blockchains. A vendor running a centralized database could also track products through a supply chain. In fact, many such systems already exist.
For example, one firm highlighted in a recent blockchain article, Provenance, proposes the tracking of fish as a use case. The firm first states the reasons for requiring a blockchain:
… There is a rallying call from customers, governments, NGOs and businesses towards the end of the supply chain for information about the origin and social standards of fish and seafood products…
The truth is that no single organization can be responsible for making data throughout a whole supply chain transparent. Third parties like NGOs or industry associations, rarely manage even one of these two aspects of transparency, and even if they could, they would become a single point of weakness. This would make them and their operations a vulnerable target for bribery, social engineering, or targeted hacking.[2]
The blockchain solution’s approach is to have multiple nodes replicating and validating the records that are linked together in the blockchain so that a change in any record would be detectible. Thus one avoids the problem of a single corrupt node providing false data.
Or does it really solve anything of importance? The nodes in the blockchain only verify that the data stored follows the agreed protocols of the blockchain system, such as the records are properly digitally signed, are of the correct form, and don’t violate other constraints. A trusted party must still decide on whose certificates to accept. In fact, Provenance also states for this use case:
The social and environmental conditions for the fishermen at the point of capture are verified through trusted local NGOs, whose audit systems validate their compliance to an external standard, resulting in their eligibility to participate in the Provenance-validated chain of custody.[3]
Supposedly we can’t trust an NGO to manage a database, but we can trust the NGO and Provenance to assure us that the producers who are registered are really following the standards and that hackers don’t steal the credentials of producers. Are site inspectors less vulnerable to bribes than database administrators? My friend who is a retired septic site inspector would disagree. He can recall many unsuccessful attempts to influence his judgment in the developed world.
Another question is not addressed. Even if the blockchain system they proposed has a lower vulnerability than alternatives, such as the “Trace My Catch” advertised by Bumblebee tuna,[4] will consumers care enough to insist on a more expensive system?
Property Titles and Ownership
Another popular use case for blockchains is tracking ownership of properties, particularly real estate. While some forms of blockchains or systems using aspects of blockchain technology may prove useful, they are more likely an incremental improvement over alternatives than a game changer.
Electronic land records and other forms of property titling is not a technology enabled by blockchain. Paperless systems were being built and implemented prior to the introduction of blockchains. A quick search can quickly show examples back to 2000 and earlier, such as that of Greene County, TN.[5]
The biggest problem with land records in the developed world is not history being altered, but fraudulent transfers, often with forged signatures, being submitted. This is acknowledged in Cook County Register of Deeds Report report on a blockchain pilot.[6] A blockchain solution alone does not solve this, unless one also insures the proper controls around submissions of deeds.[7]
While blockchains could add a means of providing a higher level of tamper resistance, it is generally at a greater cost than alternatives, a fact that is sometimes even acknowledged by blockchain consultants, such as in a recent case study from Sweden.[8] It is merely asserted that the additional protections are worth the cost. This would not be the case if less expensive technologies made corrupting the electronic records so hard that any fraud would be done by easier approaches, such as submission of invalid deeds. Furthermore, any level of resiliency and integrity could be added by increasing the number of replications and digital signatures and using some of the existing technologies that are embedded in blockchain.
Property in the Developing World
In the developing world, the problem becomes the lack of records and lack of respect for property rights. If a government wants to seize private property, it can do so with or without the blockchain registry just by filing a deed transferring the property away from the individual or declaring a deed recorded on the blockchain as no longer valid. Thus having the records on a worldwide blockchain has little advantage over other more centralized solutions.
It would not desirable to have a society where the government did not have control over deed registration and property titles were “bearer assets”. In such a system what is recorded in any blockchain or database is irrevocable truth. A victim would be unable to recover losses from a theft if the culprit used the funds to buy real property. A property owner who loses a private key would lose the ability to transfer the property, while a thief who steals the codes would permanently gain such access.
Intellectual Property
Registration of intellectual property has also been a use case. Does this really have any advantage over a central database? Once again has to trust various agents to input claim only what property they own and use the property only when they have paid the appropriate fees.
Who owns the photograph of the swallow, Andy Anderson or Jonathan Harris? A blockchain says that both are exclusive owners! Andy is able to download the loaned photo and register it, because the blockchain cannot eliminate the trust involved when data enters and exits the blockchain.
Smart Contracts
Another promise of blockchains is to enable smart contracts. The problem with this is that we are already using smart contracts. Do you have a device that allows you to drive through toll booths without stopping and have the charges debited from your account? That is a form of a smart contract. I have automated payment of electric bills, newspaper subscriptions, ride-share usage, and many other smart contracts already. Many similar innovations sold as being made possible with blockchain are already in use!
Perhaps the purists are insisting there be no written contract other than what executes on the blockchain. This position is sometimes referred to as “code is law”. The reality is that the computer programs that implement this law are even harder to read than the legalese in the traditional contracts.
Furthermore code is law, until people realize that the code does not do what they intended. Investors poured millions into a smart contract that was going to enable automatic investment in venture capital projects that participants voted for. This famous Decentralized Autonomous Organization (DAO) had a clause (referred to as a “bug”) in its smart contract that allowed someone to siphon funds into his own account. The people controlling the ethereum blockchain decided this was theft and altered the blockchain system to undo this transfer.[9]
Freedom from Government and Central Control
Blockchain solutions are often portrayed as a way of sharing data without trusting any central authority, including a government. While I have shown examples of how the blockchain system still allows breaches of trust through variations of the garbage-in — garbage-out principal, there is a more insidious misconception that blockchains eliminate the need for governance or central control.
Many blockchain implementations are private blockchains, where some organization decides who can add data to and perform various tasks on the chain. The organizer may run nodes, or induce participants to run nodes that maintain the consensus on what constitutes the approved data in the blockchain. Such organizers and the collection of nodes become the government of the blockchain.
Does this really mean authority is any more decentralized than the authority that exists when users agree to form a consortium that operates under a charter or just hire a third party provider? Instead of committee members voting or employees making decisions, those who run the nodes vote by applying their rules to accepting or rejecting transactions. The organizer exercises control by choosing the software, the participants, and the roles of each participant. One should ask whether this has a substantially different impact or just makes it harder to identify accountability.
This problem applies even to the public blockchains, including bitcoin; however in a much more obscure way. The bitcoin blockchain’s government is the collection of miners and nodes. Miners decide which transactions to include in the blocks and nodes and miners decide which blocks to accept. Currently over 70% of the mining capacity is concentrated in seven pools headquartered in China, and over 50% is concentrated in four of these.[10]
Are the leaders of these mining pools any less corruptible than the leaders of nation states, especially in places where the leaders are elected by popular vote? It is true that the leaders of the mining pools are constrained by the possibility of losing users and seeing the value of the cryptocurrency fall. While this is currently a safeguard, in a world where critical information is stored or ownership recorded on a blockchain, voting with your feet may be not much easier than leaving a country that does not respect your property rights. Furthermore, a miner’s interest may not align with the customers’ interest. We can see the workings of this governance process in the current fight over changes to the bitcoin protocol that could affect the cost and speed of approving transactions.[11]
A Digital Perpetual Motion Machine for Trust
The current enthusiasm over blockchain appears to revolve around a belief that mathematics and computing can solve the age old problem of trust. The problem is that trust ultimately extends into the physical and human world. It involves human interactions and judgments about quality, intent, abilities, and more. Perhaps the search for a computational system that uses math to eliminate the need for trust is the digital equivalent of the search for a machine that could produce power without consuming energy — the perpetual motion machine.
[1] Technically typically a private key corresponding to an appropriate public key. People don’t memorize their private keys, they gain access to them with their password.
[2] https://www.provenance.org/tracking_tuna_on_the_blockchain as of July 07, 2017
[3] https://www.provenance.org/tracking_tuna_on_the_blockchain as of July 07, 2017
[4] See http://www.bumblebee.com/tracemycatch/
[5] http://www.greenevillesun.com/news/register-of-deeds-office-now-paperless/article_121de165-8944-5a91-bbbe-c1db55b0cef9.html
[6] http://cookrecorder.com/wp-content/uploads/2016/11/Final-Report-CCRD-Blockchain-Pilot-Program-for-web.pdf see page 45 for example
[7] Some claim that blockchain solutions include features such as digital signing to prevent this; however, these features are already implemented in many existing systems.
[8] See for example https://chromaway.com/papers/Blockchain_Landregistry_Report_2017.pdf .
[9] https://www.bloomberg.com/features/2017-the-ether-thief/
[10] https://blockchain.info/pools as of July 12, 2017 . Numbers change with time.
[11] https://www.bloomberg.com/news/articles/2017-07-10/bitcoin-risks-splintering-as-civil-war-enters-critical-month
